4.3 Data and Sample
4.4.2 Annual Low-Wage Transitions
To describe low-wage transitions, we compute the extent of aggregate state depen- dence (𝐴𝑆𝐷), defined as the difference in the probabilities of low pay conditional on being initially low paid and highly paid in period 𝑡 − 1. Thus, 𝐴𝑆𝐷 is defined as:
𝐴𝑆𝐷 = 𝑃 (𝐿𝑡= 1|𝐿𝑡−1= 1) − 𝑃 (𝐿𝑡= 1|𝐿𝑡−1= 0), (4.1)
with 𝐿𝑡 = 1 and 𝐿𝑡 = 0 meaning low and high pay in year 𝑡, respectively. To
illustrate the evolution of low-wage persistence, Figure 4.2 plots 𝐴𝑆𝐷 against time. Distinguishing the pre-unification (to the left of the first vertical line), the transition (between the vertical lines) and the post-unification period (i.e. the time after mon- etary union), several noteworthy facts emerge from Figure 4.2 Panel (A). During the pre-transition period, aggregate state dependence varied around 42%. During transition, aggregate state dependence decreased markedly by more than 20% points to 24% in 1990 compared to the pre-unification period. The third part of the figure (to the right of the second vertical line) indicates a sharp rise in low-pay persistence, with aggregate state dependence increasing from 40% in 1991 to values above 70% in the late 1990s. The figure further shows that the level of persistence comes from individuals who were already low paid in t-1. Before 1990 about 5% transitioned from high pay into low pay and this number decreased after 1990 to about two percent. Figure 4.2 Panel (B) shows the evolution of aggregate state dependence by gender. The figure reveals that the sharp decline in aggregate state dependence during transition is mostly accounted for by female workers. After Reunification, there appears to be a strong convergence between male and female workers.
53In Appendix 4.D, we further validate the data by estimating conditional low-wage probabilities
given observable characteristics. The results show reasonable correlations for both males and females with respect to, e.g., education, age and occupation.
0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1 Percentage 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 Year
Agg. state dependence Low−wage | low (t−1) Low−wage | high (t−1) (A) Overall 0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1 Percentage 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 Year Males Females (B) by Gender Source: BASiD 2007, weighted sample.
Notes: Panel (A) presents aggregate state dependence and conditional probability estimates of low-wage employment given low-wage and high-wage employment in t-1 for both, males and females. The measure of aggregate state dependence is the difference between the two conditional estimates. Panel (B) presents aggregate state dependence by gender. The first vertical line marks the fall of the Berlin Wall on 89/11/09. The second line marks the monetary union on 07/01/1990. The first decile represents the low-wage threshold. Robustness checks changing the low-wage threshold to two-thirds of the median are available upon request and do not differ strongly from the results presented here.
Figure 4.2: Aggregate state dependence
The overall picture that emerges from Figure 4.2 is that aggregate state depen- dence plummeted with the beginning of a market-orientated economy. The post- unification period is characterized by a steady rise in low-pay persistence during the first years and a level-off at the end of the 1990s. Overall, the figures show that the importance of previous low-wage employment for low-pay in the current period reaches its minimum during the time of transition, albeit less pronounced for male workers.
4.4.3 Relationship between Low-Wage Employment before
and after Reunification
In what follows, we provide some descriptives on the relationship between the inci- dence of low-wage employment before and after Reunification. As a first measure of pre-unification low-wage experiences, we count the number of years in which an individual’s earnings fell below the first decile between 1980 and 1989. In our sample there are 1074 (49%) male and 637 (27%) female workers who spent at least one year below the first decile of the wage distribution during 1980 and 1989. To obtain sufficient observations per cell, we construct three GDR low-wage experience cate- gories. The first one corresponds to individuals whose (average monthly) earnings
within a given year never fell short of the GDR low-wage threshold. The second 0 5 10 15 20 25 30
Share of Low−wage Individuals
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999
Sample Year
Overall 1−3 GDR low−wage years
> 3 GDR low−wage years (A) Males 0 5 10 15 20 25 30 35 40 45 50
Share of Low−wage Individuals
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999
Sample Year
Overall 1−3 GDR low−wage years
> 3 GDR low−wage years
(B) Females Source: BASiD 2007, weighted sample.
Figure 4.3: Percentage of low-wage employment conditional on the number of GDR years below the first decile before Reunification
(third) group experienced one to three (more than three) years of GDR low-wage employment. Among those with at least one year below the first decile, 90% of males and 66% of females experienced between one to three years of low wages. Figure 4.3 reports the descriptive statistics of the probability of being in low-wage employment over the time period between 1990 and 1999 for males and females. Moreover, we plot the same probability conditional on the variable number of GDR years below the first decile. From the figure, we first observe that the low-wage sector has been increasing over time for both males and females. Moreover, experiencing a low-wage period between one to three years during 1980 and 1989 raises the probability of belonging to the low-wage sector after Reunification only slightly for male workers, whereas the difference becomes more pronounced for females. The probability of being low paid after Reunification increases further, if the number of pre-unification low-wage years exceeded three years. Interestingly, from both figures we observe that, conditional on more than three years pre-unification low-wage employment, the probability of being low paid does not exhibit an increasing trend as in the pooled sample.